High accuracy SPICE behavioral macromodeling of insulated gate bipolar transistor (IGBT)

This paper presents a new behavioral IGBT macromodel, that uses the enhanced capabilities of the nonlinear controlled sources implemented in modern SPICE like simulators. It describes the device's internal static equations directly with "in line equation" controlled sources and the nonlinear voltage dependent gate capacitances are piece-wise-linear approximated with "look-up table" controlled sources. The temperature influence on model parameters and optionally the self-heating and thermal coupling with adjacent devices are also included. The proposed model considers the current modulated base resistance and the forward and reverse device breakdown. The parameters extraction algorithm was greatly simplified, as the classic curve fitting was replaced by a direct specification of data-sheets characteristics as model parameters. This new behavioral macromodel assures a higher accuracy of static and dynamic IGBTs' description, with no convergence problems and with a reasonable analysis time.